Mutual Diffusion Coefficients of Aqueous KCl at High Pressures

Sep 26, 2011 - South Kensington Campus, SW7 2AZ, London, United Kingdom. ‡ ... of Bucharest, 1-7 Gh. Polizu Street, S1, 011061 Bucharest, Romania...
1 downloads 0 Views 1MB Size
ARTICLE pubs.acs.org/jced

Mutual Diffusion Coefficients of Aqueous KCl at High Pressures Measured by the Taylor Dispersion Method Catinca Secuianu,*,†,‡ Geoffrey C. Maitland,† J. P. Martin Trusler,† and William A. Wakeham† †

Qatar Carbonates and Carbon Storage Research Centre (QCCSRC), Department of Chemical Engineering, Imperial College London, South Kensington Campus, SW7 2AZ, London, United Kingdom ‡ Department of Applied Physical Chemistry and Electrochemistry, Faculty of Applied Chemistry and Materials Science, “Politehnica” University of Bucharest, 1-7 Gh. Polizu Street, S1, 011061 Bucharest, Romania ABSTRACT: Diffusion coefficients of potassium chloride in potassium chloride solutions of molality (0, 1, 2.5, and 4.5) mol 3 kg1 were measured at temperatures of (298.15, 323.15, 348.15, 373.15, and 423.15) K and at pressures from (0.1 to 69) MPa using the Taylor dispersion technique. The results have an overall estimated expanded relative uncertainty of 1.6 % with a coverage factor k = 2. When compared with standard reference data for the diffusion coefficient of KCl in KCl solutions at T = 298.15 K and p = 0.l MPa, a relative deviation of 0.5 % is observed. The effect of pressure was found to be very small. A satisfactory multiparameter correlation of all of the data in terms of pressure, temperature, and molality was obtained; this exhibits an absolute average relative deviation of 1.2 % in comparison with our data. The model also provides a good account of the majority of the data available in the literature.

’ INTRODUCTION The thermophysical properties of systems comprising water, salts, hydrocarbons, and/or carbon dioxide are important factors in the design, optimization, and control of water and gas injection processes in both petroleum reservoirs and saline aquifers. Carbonate reservoirs are especially interesting as strong chemical interactions between reservoir fluids and minerals are possible. The present research contributes to a program of measurements and modeling of the physical and chemical properties of systems comprising carbon dioxide, hydrocarbons, brines, and minerals at reservoir conditions that is being undertaken as part of a multidisciplinary research center.1 The research is directed toward an improved understanding of the detailed physical and chemical processes occurring at the pore- and fracture-scale in carbonate reservoirs during water and gas injection processes, including both geological storage of CO2 and CO2-enhanced hydrocarbon production. Diffusion is one of the properties that are of interest, especially at the pore-scale. To measure this property for systems containing carbon dioxide and/or hydrocarbons and/or brines at reservoir conditions, a new apparatus has been built. The method used is the chromatographic peak broadening technique which is based on the fundamental work of Taylor,2,3 as extended by Aris,4 which involves the dispersion of a solute injected into the steadystate laminar flow of a mobile solvent phase passing through a long tube of uniform diameter. Due to the combined effects of convective flow and molecular diffusion, an initially sharp pulse develops a Gaussian distribution. The temporal variance of this Gaussian distribution is dependent on both the average flow velocity and the molecular diffusivity. At the end of the diffusion tube, the concentration is measured as a function of time. In the present work, this was accomplished by means of a chromatographic refractive index detector. As built, the apparatus is capable of operating with the diffusion column at temperatures of up to 473 K and pressures up to 70 MPa. In this work, we begin by studying a simple electrolyte solution, aqueous potassium chloride, with two r 2011 American Chemical Society

objectives. The first was to validate the new experimental apparatus by means of measurements at ambient pressure and T = 298.15 K where the diffusion coefficient of KCl in pure water is well-known. The second was to gather data for diffusion of KCl in KCl solutions over extended ranges of temperature, pressure, and molality. The Taylor dispersion technique for measuring diffusion coefficients in liquids was introduced by Pratt and Wakeham5 and Grushka and Kikta.6 The method has a number of advantages including relatively rapid measurements, ease of automation, use of standard high-performance liquid chromatography (HPLC) components, and especially, a fully developed working equation that permits absolute measurements to be made. Since these early works, the method has been widely used for the measurement of diffusion coefficients in a wide range of fluids and conditions,711 especially organic mixtures,12,13 and in supercritical fluids.1420 Tominaga et al.,21 Leaist,22 and Castillo and Garza23 showed that this method can be applied successfully for measuring the diffusion coefficient of electrolytes at atmospheric pressure. An extensive literature review revealed that numerous papers2372 report diffusion coefficients for aqueous potassium chloride solutions, but most of them are measured at ambient conditions and for dilute solutions. The majority of the studies relating to diffusion in aqueous solutions of electrolytes are concerned with the development of reliable techniques for the measurement of the concentration dependence.23 In addition, according to Lobo et al.,73 there are few cases in which the results from different researchers can be compared directly. However, the diffusion coefficients of aqueous potassium chloride solutions Special Issue: Kenneth N. Marsh Festschrift Received: July 28, 2011 Accepted: September 9, 2011 Published: September 26, 2011 4840

dx.doi.org/10.1021/je200808q | J. Chem. Eng. Data 2011, 56, 4840–4848

Journal of Chemical & Engineering Data

Figure 1. Schematic diagram of the Taylor dispersion experimental apparatus: S, solvent bottle; D, degasser; P, pump; V, six ports valve; W, waste; B, oil bath; DT, diffusion tubing; RT, restrictor tubing; TR, thermometer readout; RID, refractive index detector; PC, computer.

at T = 298.15 K and at concentrations of up to 1 mol 3 L1 are recommended50,74,75 as a reference standard for the liquid phase. In this work, we measured the diffusion coefficients for KCl in aqueous KCl solutions at molalities of (0, 1, 2.5, and 4.5) mol 3 kg1, pressures up to 69 MPa, and temperatures ranging from (298.15 to 423.15) K. As far as we are aware, this is the first study of the KCl(aq) system to address high pressures. The ability to work at an elevated pressure also facilitated measurements at temperatures above the normal boiling point of the solution, where no previous data have been reported.

’ EXPERIMENTAL SECTION Materials. Potassium chloride was supplied by Sigma-Aldrich with a claimed mass fraction purity >0.990; the KCl was dried in an oven at T = 423 K but otherwise used without further purification. Ultrapure deionized water with an electrical resistivity of > 18 MΩ 3 cm was obtained from a Millipore system (Direct-Q, Millipore UK, Ltd.). Apparatus and Procedure. Figure 1 is a schematic diagram of the Taylor dispersion apparatus used in this work comprising three modules: a solvent delivery system with solute injection valve, a thermostatic oil bath housing the diffusion capillary, and a differential refractive-index detector. The solvent was drawn through a filter from the supply bottle via an inline degasser (Knauer, Model A5328, Germany) to charge the 100 mL capacity syringe pump (Teledyne ISCO 100DM Hastelloy, USA), which was kept at a constant temperature of 293 K by circulating water supplied by a chiller unit (Huber, model Minichiller-NR, Germany) through a cooling jacket fitted to the syringe. The solvent was delivered at a constant flow rate of between (0.05 and 0.425) mL 3 min1, through a six-port injection valve (VICI Cheminert, model C72H-1696D, Switzerland) fitted with a 5 μL sample loop, into the diffusion capillary. The latter was maintained at the desired temperature by immersion in a thermostatic oil bath (Fluke Hart Scientific, model 6022, USA) controlled to within 0.01 K. The outflow from the column passed through a short restrictor tube and into the refractive index detector (Agilent 1200 Series, model G1362A, USA), from which it passed to waste. Preliminary tests showed that even a passivated stainless steel diffusion tube was not suitable for KCl solutions, due to pitting

ARTICLE

corrosion. Therefore, a Hastelloy C-276 tube of nominal external radius 0.8 mm (Vindum Engineering Inc., USA) was used. The length of the tube, L = (4.52 ( 0.01) m, was measured with an ordinary tape measure. The internal radius of the tube, R = (0.541 ( 0.001) mm, was determined by weighing the tube, both empty and after filling with pure water, on an analytical balance having a resolution of 0.1 mg. For the installation in the oil bath, the tube was coiled on a drum of radius of 0.109 m. The temperature of the oil bath was measured with a secondary-standard platinum resistance thermometer (Fluke Hart Scientific model 5615) and readout unit (Fluke Hart Scientific, model 1502A). The thermometer was calibrated on ITS-90 at the temperature of the triple point of water and by comparison in a constant temperature bath with a standard platinum resistance thermometer at nominal temperatures of (323, 373, 423, and 473) K. The expanded uncertainty of the temperature measurements was 0.02 K. The solvent delivery pressure was measured at the outflow of the pump by means of the pressure transducer integrated into the pump. According to the manufacturer, the relative uncertainty of the pressure was 0.5 % of reading. The refractive index detector used in this work was limited to a maximum operating pressure of 0.5 MPa, and as the outlet was open to the atmosphere, it was in practice operated close to ambient pressure. To permit elevated pressures upstream, a selection of restrictor tubes were inserted between the column and the detector with a length and diameter chosen to provide the desired back pressure. The restrictor tubes were either a Siltek (deactivated stainless steel) capillary of length 0.58 m and internal radius 0.127 mm, for low-pressure operation (p < 1 MPa) or PEEK clad fused silica tubes with lengths between (0.05 and 0.5) m and an internal radius of 12.5 μm for higher pressures. With a given restrictor tube fitted, the column pressure could be varied by changing the solvent flow rate within the range stated above. The flow rate accuracy of the pump was checked by weighing the outflow under steady-flow conditions for a precisely timed period of about 1 h. This was repeated for three different flow rates spanning the normal experimental range, and the results were all within ( 0.35 % of the set flow rate. The experiment was operated under computer control using a program written in Agilent VEE that logged the experimental temperature, actuated the injection valve, and collected readings every 5 s from the refractive index detector. The pump was operated under manual control at the chosen flow rate and the pressure under steady-state flow was recorded before, during, and after the passage of a solute injection through the column. Once steady-state flow and thermal equilibrium were established, the measurements were started. The solute solution was loaded into the sample loop and injected into the solvent stream at time t = 0. At least three solute injections were made per run, and they were spaced at (15 to 30) min intervals to avoid any overlap of peaks. Solutions were carefully prepared by dissolving weighed amounts of dried KCl in a known amount of the mobile phase, typically to produce a solution with a molality exceeding that of the solvent by 0.1 mol 3 kg1. As discussed in the literature,76 the solute concentration (mass per unit volume of solution) averaged over the cross section at the end of the diffusion tube is given by: " # m ðL  vtÞ2 exp  cðtÞ ¼ ð1Þ 4Kt πR 2 ð4πKtÞ1=2 4841

dx.doi.org/10.1021/je200808q |J. Chem. Eng. Data 2011, 56, 4840–4848

Journal of Chemical & Engineering Data

ARTICLE

Figure 2. Differential refractive index signal s(t) = Δn for the measurement of the diffusion coefficient of KCl in water at T = 298.15 K and p = 0.1 MPa: O, experimental data; solid line, fit with eq 3.

In eq 1, m is the mass of solute injected, R is the internal radius of the column, L is the length of the column, v is the axial velocity of the solvent averaged over the cross section of the tube, t is the time, and K is the dispersion coefficient which is related to the binary diffusion coefficient, D, by4 K ¼D þ

R 2 v2 48D

ð2Þ

Wakeham and co-workers7779 discussed the constraints that must be placed on the experimental conditions to satisfy the assumptions made in the derivation of eq 1, as well as the corrections required. The dimensions of the diffusion tubing, the coil radius, and the flow rates were therefore chosen to meet the criteria discussed in those studies. Thus, in our experiments, the flow was always laminar as the Reynolds number, Re, was in the range 1.3 < Re < 24. To ensure that secondary flow effects are negligible, the criterion De2Sc < 20 must be met, where De = Re(R/Rcoil)1/2 is the Dean number, Sc = η/(FD) is the Schmidt number, Rcoil is the coil radius, η is the solvent viscosity, and F is the solvent density. In the present work, this criterion was easily exceeded as the experiment was operated such that 0.01 < De2Sc < 0.5. Finally, the criterion for the concentration profile resulting from dispersion of the initial solute pulse to become spatially Gaussian (also known as the LevenspielSmith80 criterion) is K/(vL) < 0.01; in our experiments this was always fulfilled as 3.8 3 104 < K/(vL) < 7.6 3 103. The diffusion coefficient can be calculated from the measured concentrationtime curves in different ways:81 the heightarea method, moments calculations, or by fitting the parameters of the following equation to the experimental data: sðtÞ ¼ a þ bt þ αcðtÞ

ð3Þ

Here, s(t) is the signal at the detector, parameters a and b compensate for baseline drift, assumed to be linear in time, α is the detector sensitivity, and c(t) is given by eq 2. The latter method is generally the more accurate and therefore preferred in this work. Figure 2 illustrates an example set of data together with the fitting curve. The average absolute deviations between the

Figure 3. Relative deviations δΔn/n between the experimental differential refractive index signal s(t) = Δn and the fit with eq 3 as a function of time t for KCl in water at T = 298.15 K and p = 0.1 MPa.

experimental data and the fitting curve are approximately 0.5 % of the maximum signal (Figure 3). The detector used measures not concentration directly but the difference Δn between the refractive index of the eluent at time t and that of the solvent. There is therefore an assumption that, for small changes in concentration c, Δn is proportional to Δc. In the present case, Δc was of order 0.1 g 3 dm3 which may be considered sufficiently small so as to ensure linearity. One of the difficulties encountered with diffusion coefficient measurements by the Taylor method is the influence of injection, detection, and connecting tube volumes.77 An analysis of the effects of these hold-up volumes to either side of the diffusion tubing on the interpretation of the final signal was carried out. This showed that the effect of tubing sections of radius Ri and length Li inserted before or after the column combine to introduce excess dispersion that is equivalent to an additional length δL of the main column, where δL ¼

∑i fDðTÞ=DðTi ÞgðRi =RÞ4 Li

ð4Þ

Here Ti is the temperature for tubing section i. Accordingly, the effective column length used in the analysis was increased by δL. In the configuration used for the present work, there were three significant sections of additional tubing: between the injection valve and the column (R1 = 0.39 mm, L1 = 0.3 m), between the column and the inlet of the detector (R2 = 0.13 mm, L2 = 0.6 m), and finally the internal flow path in the detector itself (R3 = 0.22 mm, L3 = 0.4 m). For example, with D(Ti) = 2 3 109 m2 3 s1 in each section this led to δL = 0.096 m or about 2 % of L. These corrections could be made smaller by reducing the internal radius of these tubing sections, especially the first, or by increasing the length of the main column. The restrictor tubes added to obtain high column pressures had negligible additional effect on the dispersion due to their very small internal radii. In analyzing the results, we obtained v from the programmed flow rate. Inevitably, the actual flow rate differed slightly from the programmed value and, to account for this in the analysis, the time origin for each peak was an adjustable parameter. However, 4842

dx.doi.org/10.1021/je200808q |J. Chem. Eng. Data 2011, 56, 4840–4848

Journal of Chemical & Engineering Data

ARTICLE

Table 1. Experimental Values of Diffusion Coefficients D at Temperature T, Pressure p, and Molality m for KCl in KCl(aq)a T K

p

D 3 109

MPa

m2 3 s1 1

m/mol 3 kg

p

D 3 109

MPa

m2 3 s1

T K

Table 1. Continued T K

0.16

1.945

348.15

29.27

4.441

298.15

14.99

1.942

348.15

46.78

4.495

298.15

30.92

1.947

348.15

66.70

4.522

298.15

48.17

1.954

373.15

0.82

6.107

298.15 323.15

62.94 0.64

1.968 3.022

373.15 373.15

24.18 43.13

6.145 6.112

323.15

15.46

3.016

373.15

64.28

6.182

323.15

33.58

3.085

423.15

3.08

10.263

323.15

68.56

3.093

423.15

18.51

10.271

348.15

0.65

4.481

423.15

33.58

10.329

348.15

15.72

4.488

423.15

63.30

10.310

m/mol 3 kg1 = 1.0 ( 0.0001 298.15 298.15

0.24 10.17

1.892 1.907

348.15 348.15

39.76 41.21

4.232 4.245

298.15

28.21

1.918

348.15

41.98

4.249

298.15

38.63

1.925

348.15

45.33

4.263

298.15

45.59

1.947

348.15

62.42

4.243

298.15

59.03

1.958

348.15

64.02

4.247

298.15

67.07

1.978

348.15

67.09

4.250

323.15

0.61

3.034

373.15

0.82

5.942

323.15 323.15

15.30 43.40

3.062 3.084

373.15 373.15

23.41 42.09

5.932 5.936

323.15

52.06

3.095

373.15

64.21

5.973

323.15

68.97

3.097

423.15

3.41

9.432

348.15

0.66

4.229

423.15

15.13

9.446

348.15

25.82

4.246

423.15

34.72

9.455

348.15

31.12

4.257

423.15

61.62

9.479

348.15

35.87

4.241

423.15

65.64

9.486

1

m/mol 3 kg

D 3 109

MPa

m2 3 s1

T K

p

D 3 109

MPa

m2 3 s1

m/mol 3 kg1 = 4.5 ( 0.0001

= 0.0 (infinite dilution)

298.15

p

298.15

0.48

2.172

348.15

65.91

298.15

27.06

2.144

373.15

0.74

4.628 6.251

298.15

43.80

2.171

373.15

17.78

6.222

298.15

66.02

2.184

373.15

41.28

6.239

323.15

0.50

3.335

373.15

47.90

6.209

323.15

3.95

3.324

373.15

54.29

6.244

323.15

31.37

3.347

373.15

54.41

6.240

323.15 323.15

49.57 62.69

3.285 3.352

373.15 423.15

65.70 13.02

6.274 9.265

348.15

0.73

4.645

423.15

40.54

9.294

348.15

12.83

4.587

423.15

68.93

9.301

348.15

32.97

4.619

a Standard uncertainties u are u(T) = 0.02 K, u(p) = 0.005 3 p, u(m) = 0.0001 mol 3 kg1, and the combined expanded uncertainty Uc(D) = 0.0035 3 D (level of confidence = 0.95).

= 2.5 ( 0.0001

298.15

0.29

2.018

298.15

37.39

2.038

298.15

2.29

2.023

298.15

47.43

2.043

298.15

4.50

2.039

298.15

64.97

2.069

298.15

24.92

2.037

298.15

66.17

2.079

298.15

32.60

2.047

298.15

67.63

2.080

323.15

0.58

3.152

373.15

24.36

5.940

323.15

40.89

3.122

373.15

31.32

5.886

323.15

45.21

3.177

373.15

47.55

5.984

323.15

47.48

3.173

373.15

57.26

5.857

323.15

60.49

3.118

373.15

59.59

5.955

323.15

64.86

3.152

373.15

63.05

5.961

348.15

0.70

4.522

423.15

4.44

9.685

348.15

18.05

4.548

423.15

15.18

9.673

348.15

29.96

4.573

423.15

15.26

9.675

348.15

45.15

4.580

423.15

34.63

9.698

348.15

68.90

4.594

423.15

34.63

9.698

373.15 373.15

0.94 17.14

5.920 5.915

423.15 423.15

56.03 65.05

9.702 9.716

Figure 4. Diffusion coefficients of KCl in KCl(aq) as a function of molality m at T = 298.15 K ([, this work; , Woolf and Tilley;50 /, Pinto and Graham;62 4, Lobo et al.71) and at T = 323.15 K (9, this work; O, Lobo;64 +, Fell and Hutchison.54).

fitted value was normally within 5 s of the actual time of the solute injection, which may be compared to a typical elution time of 2000 s. A further detail, to be considered when the experimental temperature differs from that in the syringe pump, is the effect of thermal expansion on the volumetric flow rate in the column. Corrections for this small effect were included.

’ RESULTS AND DISCUSSION The measured diffusion coefficients of potassium chloride in pure water (infinitesimal concentration) and in potassium chloride solutions, together with the temperatures, pressures, and molalities at which the experiments were carried out, are given in Table 1. The values of the diffusion coefficients are averages of at 4843

dx.doi.org/10.1021/je200808q |J. Chem. Eng. Data 2011, 56, 4840–4848

Journal of Chemical & Engineering Data

ARTICLE

Table 2. Parameters of Equations 7 to 9 a0

a1

a2

a3

a4

b0

b1

c0

c1

15.5

0.315

0.085

0.684 3 102

0.132 3 103

910.2

62.6

134343

2007.2

Figure 5. Diffusion coefficients of KCl in KCl(aq) as a function of pressure at (a) infinite dilution, (b) m = 1.0 mol 3 kg1, (c) m = 2.5 mol 3 kg1, and (d) m = 4.5 mol 3 kg1: [, T = 298.15 K; 9, T = 323.15 K; 2, T = 348.15 K; b, T = 373.15 K; /, T = 423.15 K; solid line, eqs 6 to 9.

least three measurements, but typically five. The reproducibility of the results is generally good as the relative standard deviation ur(K) of the dispersion coefficient was 0.35 % on average, and it never exceeded 1.3 %. The estimated standard relative uncertainty of D may be obtained from consideration of eq 2, in which the first term was always negligible. Thus ur 2 ðDÞ ¼ 4u2r ðRÞ þ 4ur 2 ðvÞ þ ur 2 ðKÞ

ð5Þ

where ur(R), ur(v), and ur(K) are the standard relative uncertainties of the column radius, flow velocity, and the dispersion coefficient. In the present case, ur(R) = 2.2 3 103, ur(v) = 3 3 103, and ur(K) = 3.5 3 103, the latter estimated from the standard

deviation of K in repeated injections. The value of ur(v) was the flowrate uncertainty specified by the pump manufacturer, rather than the smaller value implied by our three check measurements. Additional errors propagated from the uncertainties in temperature, pressure, and molality were negligible, and thus the final estimated relative expanded uncertainty of D was 1.6 % with a coverage factor k = 2. Figure 4 shows a detailed comparison of our results at T = (298.15 and 323.15) K for p < 1 MPa with existing atmosphericpressure literature data. It can be seen that D remains constant at low molalities or even decreases slightly, possibly due to increased electrostatic interactions as molality increases and mean ionic separation decreases. Above m ≈ 0.7 mol 3 kg1, the diffusion coefficient increases gently with molality due to increased screening of the electrostatic interactions. These changing electrostatics 4844

dx.doi.org/10.1021/je200808q |J. Chem. Eng. Data 2011, 56, 4840–4848

Journal of Chemical & Engineering Data

ARTICLE

Table 3. Average Absolute Relative Deviation ΔAAD and Maximum Absolute Relative Deviation ΔMAD from Equations 6 to 9 Calculated for All Available Literature Data at Temperature T, Molality m, and Atmospheric Pressure T K

Figure 6. Difference of δD diffusion coefficients D of KCl in KCl(aq) from the eqs 6 to 9 as a function of (a) temperature T, (b) pressure p, and (c) molality m.

will affect both the direct ionion interactions and their interaction with the water solvent molecules, as reflected in the effective hydrated ionic radius. Our data at (infinitesimal concentration, 1.0, 2.5, and 4.5) mol 3 kg1 are in good agreement with all sources, except Fell and Hutchison54 who measured D at very low concentrations. The relative difference between the diffusion coefficient measured in this work (1.8580 3 109 m2 3 s1) and that recommended as a standard reference value50 (1.8585 3 109 m2 3 s1) at T = 298.15 K, p = 0.1 MPa, and concentration 1 mol 3 L1 is less than 0.5 %, which confirms the reliability of our apparatus and procedure. All experimental data obtained in this work were fitted with a multiple parameter model of the form lnðD=m2 3 s1 Þ ¼ A þ BðK=TÞ þ CðK=TÞ2 1 1=2

A ¼ a0 þ a1 ðm=mol 3 kg Þ 1 2

ð6Þ

þ a2 ðm=mol 3 kg1 Þ

þ a3 ðm=mol 3 kg Þ þ a4 ðp=MPaÞ

ð7Þ

B ¼ b0 þ b1 ðm=mol 3 kg1 Þ1=2

ð8Þ

C ¼ c0 þ c1 ðm=mol 3 kg1 Þ

ð9Þ 4845

m range mol 3 kg1

Nexpa

102 3 ΔAAD

102 3 ΔMAD

ref

273.15

0.09843.0233

11

15.5

17.0

64

277.15

0.01670.5691

5

13.9

15.1

36

288.15

1.00004.0000

4

3.8

5.6

45

291.15

0.05024.5630

13

6.0

10.2

29

291.15

0.06413.7355

16

4.0

5.1

64

291.65

0.05024.5890

48

6.2

10.6

24

291.65

0.00002.0000

8

5.1

8.1

27

293.15

0.00130.0112

5

0.9

2.0

34

293.15

0.06735.1563

8

6.1

8.2

64

298.15

0.02010.5090

3

2.0

3.7

25

298.15

0.00002.1346

13

2.5

12.8

26

298.15

0.00250.0050

4

0.8

1.8

31

298.15

0.10060.0025

7

1.4

3.1

32

298.15

0.00130.0098

6

1.1

2.0

33

298.15

0.00130.5376

15

1.4

2.5

35

298.15

0.01000.5090

7

1.6

3.5

38

298.15

0.00003.9251

13

1.7

3.3

39

298.15

0.00104.4337

35

3.1

8.7

40

298.15

0.19500.2018

2

2.5

2.5

42

298.15

0.50904.4337

2

2.0

3.0

44

298.15

0.12580.5091

8

2.3

2.5

46

298.15

1.00004.0000

4

0.9

1.6

45

298.15

0.89660.9181

2

3.8

3.8

52

298.15

0.20960.2627

2

2.2

2.3

53

298.15 298.15

0.20183.3084 0.00190.5376

4 8

1.6 3.5

2.8 6.2

61 65

298.15

0.01000.5090

5

5.8

13.9

67

298.15

0.01364.6227

18

1.2

2.4

59

298.15

0.41450.5249

2

2.5

2.8

53

298.15

0.50903.3084

5

1.2

2.0

51

298.15

0.00013.3084

15

1.5

2.8

49

298.15

0.05020.5090

5

1.3

2.6

64

298.15 298.15

0.10064.4347 0.00120.5376

11 15

1.3 1.4

2.3 2.5

37 34

298.15

0.00001.0340

4

1.4

3.2

30

298.15

0.10062.1346

6

3.4

12.7

29

298.15

0.00050.0100

6

1.3

2.3

31

298.15

0.00004.0000

7

1.5

3.2

62

298.15

0.00101.0327

11

1.9

3.2

71

298.15

0.00101.0339

28

1.7

3.9

50

298.15

0.01002.0000

9

2.6

5.1

27

298.15

0.0049

1

1.4

1.4

23

303.15

0.0049

1

3.8

3.8

23

303.15

0.00320.0123

4

1.8

2.4

34

308.15

0.0049

1

6.1

6.1

23

308.15

1.00004.0000

4

7.9

10.5

45

308.15 308.15

0.05973.9767 0.10063.3084

16 6

1.0 1.9

4.4 3.4

64 47

313.15

0.0049

1

6.1

6.1

23

dx.doi.org/10.1021/je200808q |J. Chem. Eng. Data 2011, 56, 4840–4848

Journal of Chemical & Engineering Data

ARTICLE

Table 3. Continued T

m range mol 3 kg1

K

a

Nexpa

102 3 ΔAAD

102 3 ΔMAD

ref

313.15 318.15

0.00170.0224 0.0049

16 1

2.7 3.5

7.2 3.5

54 23

318.15

1.00004.0000

4

2.8

3.1

45

318.15

0.10063.3084

6

2.4

2.8

47

323.15

0.00230.0393

31

4.8

7.3

54

323.15

0.09894.2832

19

3.1

4.0

64

333.15

0.00260.0378

32

4.9

7.3

54

343.15

0.00300.0369

9

5.0

6.3

54

353.15

0.00460.0101

4

7.3

7.5

54

Number of experimental points.

ΔAAD ¼ N

N

∑ jðDi, exp  Di, calc Þ=Di, exp j

ð10Þ

i¼1

and by the maximum absolute relative deviation, ΔMAD ¼ Max i jðDi, exp  Di, calc Þ=Di, exp j

ð11Þ

where N is the number of experimental points (N = 113) and subscripts “exp” and “calc” refer to experimental and calculated values of D at the ith state point. For the present data, ΔAAD = 1.2 % and ΔMAD = 3.8 %. The experimental diffusion coefficients obtained in this study are plotted as a function of pressure at constant temperature and molality in Figure 5 together with isotherms computed from eqs 6 to 9. It is clear that the dependence upon pressure is very weak, presumably as a consequence of the low compressibility of these aqueous solutions. However, the parameter a4, which accounts for the effect of pressure in our correlation, was found to make a significant improvement in the fit and so was retained. The relative deviation of the data from the correlation is shown in Figure 6 as functions of temperature, pressure, and molality; it can be seen that most are < 2.0 % and that there are no obvious systematic trends. The predictions of the empirical fit were tested by a comparison with all available literature data (some 556 experimental points, all at ambient pressure), and the deviations for each literature source are summarized statistically in Table 3. Considering all literature sources, ΔAAD = 3.5 %. Where necessary, concentration units were converted from molarity C to molality m, using the equation m ¼ C=ðF  CM2 Þ

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Tel.: +44 20 7594 1341. Fax: +44 20 7594 5700. Funding Sources

where ai, bi, and ci are parameters, the values of which are given in Table 2. The goodness of fit was measured by the average absolute relative deviation (ΔAAD), 1

of molality, temperature, and pressure by means of the Taylor dispersion method. The new dispersion apparatus was validated by comparing the results with very accurate literature data for aqueous KCl at ambient pressure, T = 298.15 K, and a concentration of 1 mol 3 L1. The new data significantly extend the range of temperature over which diffusion coefficients of KCl in KCl(aq) are known. The effect of pressure on D was found to be very small.

ð12Þ

where F is the solution density and M2 is the molar mass of the solute. Density and viscosity values necessary for the calculation of the m, Re, and Sc criteria were taken from the NIST82 and DETHERM 200783 databases.

’ CONCLUSIONS Diffusion coefficients of potassium chloride in water and potassium chloride aqueous solutions were measured as function

This work was a part of the activities of the Qatar Carbonates & Carbon Storage Research Centre (QCCSRC). We gratefully acknowledge the funding of QCCSRC provided jointly by Qatar Petroleum, Shell, and the Qatar Science & Technology Park.

’ ACKNOWLEDGMENT We gratefully acknowledge QCCSRC and Qatar Petroleum, Shell, and the Qatar Science & Technology Park for their permission to publish this research. C.S. would like to thank Prof. Amparo Galindo for helpful discussions. ’ REFERENCES (1) http://www3.imperial.ac.uk/qatarcarbonatesandcarbonstorage (accessed July 28, 2011). (2) Taylor, G. Dispersion of soluble matter in solvent flowing slowly through a tube. Proc. R. Soc. London 1953, A219, 186–203. (3) Taylor, G. Conditions under which dispersion of a solute in a stream of solvent can be used to measure molecular diffusion. Proc. R. Soc. London 1954, A225, 473–477. (4) Aris, R. On the dispersion of a solute in a fluid flowing through a tube. Proc. R. Soc. London 1956, A235, 67–77. (5) Pratt, K. C.; Wakeham, W. A. The mutual diffusion coefficient of ethanol-water mixtures: determination by a rapid, new method. Proc. R. Soc. London 1974, A336, 393–406. (6) Grushka, E.; Kikta, E. J. Extension of the chromatographic broadening method of measuring diffusion coefficients to liquid systems. I. Diffusion coefficients of some alkylbenzenes in chloroform. J. Phys. Chem. 1974, 78, 2297–2301. (7) Weinheimer, R. M.; Fennell Evans, D.; Cussler, E. L. Diffusion in surfactant solutions. J. Colloid Interface Sci. 1981, 80, 357–368. (8) Atwood, J. G.; Goldstein, J. Measurements of diffusion coefficients in liquids at atmospheric and elevated pressure by the chromatographic broadening technique. J. Phys. Chem. 1984, 88, 1875–1885. (9) Chen, H.-C.; Chen, S.-H. Taylor dispersion measurement of the diffusivities of polymethylenes in dilute solutions. Chem. Eng. Sci. 1985, 40, 521–526. (10) van de Ven-Lucassen, I. M. J. J.; Kieviet, F. G.; Kerkhof, P. J. A. M. Fast and Convenient Implementation of the Taylor Dispersion Method. J. Chem. Eng. Data 1995, 40, 407–411. (11) Sarraute, S.; Costa Gomes, M. F.; Padua, A. A. H. Diffusion coefficients of 1-alkyl-3-methylimidazolium ionic liquids in water, methanol, and acetonitrile at infinite dilution. J. Chem. Eng. Data 2009, 54, 2389–2394. (12) Funazukuri, T.; Nishimoto, N.; Wakao, N. Binary diffusion coefficients of organic compounds in hexane, dodecane, and cyclohexane at 303.2333.2 K and 16.0 MPa. J. Chem. Eng. Data 1994, 39, 911–915. (13) Higashi, H.; Iwai, Y.; Arai, Y. Review. Solubilities and diffusion coefficients of high boiling compounds in supercritical carbon dioxide. Chem. Eng. Sci. 2001, 56, 3027–3044. 4846

dx.doi.org/10.1021/je200808q |J. Chem. Eng. Data 2011, 56, 4840–4848

Journal of Chemical & Engineering Data (14) Levelt Sengers, J. M. H.; Deiters, U. K.; Klask, U.; Swidersky, P.; Schneider, G. M. Application of the Taylor dispersion method in supercritical fluids. Int. J. Thermophys. 1993, 14, 893–922. (15) Bruno, T. J. Measurement of diffusion in fluid systems: applications to the supercritical fluid region. J. Thermophys. Heat Trans. 1994, 8, 329–333. (16) Catchpole, O. J.; King, M. B. Measurement and correlation of binary diffusion coefficients in near critical fluids. Ind. Eng. Chem. Res. 1994, 33, 1828–1837. (17) Silva, C. M.; Macedo, E. A. Diffusion coefficients of ethers in supercritical carbon dioxide. Ind. Eng. Chem. Res. 1998, 37, 1490–1498. (18) Suarez, J. J.; Medina, I.; Bueno, J. L. Review. Diffusion coefficients in supercritical fluids: available data and graphical correlations. Fluid Phase Equilib. 1998, 153, 167–212. (19) Karaiskakis, G.; Gavril, D. Review. Determination of diffusion coefficients by gas chromatography. J. Chromatogr., A 2004, 1037, 147–189. (20) Funazukuri, T.; Kong, C. Y.; Kagei, S. Binary diffusion coefficients in supercritical fluids: recent progress in measurements and correlations for binary diffusion coefficients. J. Supercrit. Fluids 2006, 38, 201–210. (21) Tominaga, T.; Matsumoto, S.; Koshiba, T.; Yamamoto, Y. Tracer diffusion of the tris(1,10-phenanthroline)iron(II) cation in aqueous salt solutions. Effect of hydrophobic interactions. J. Chem. Soc., Faraday Trans. 1 1988, 84, 4261–4266. (22) Leaist, D. G. Tracer ionic diffusion coefficients from liquid chromatography. Electrochim. Acta 1991, 36, 309–313. (23) Castillo, R.; Garza, C. Temperature dependence of the mutual diffusion coefficients in aqueous solutions of alkali metal chlorides. Int. J. Thermophys. 1993, 14, 1145–1152. (24) Clack, B. W. On the study of diffusion in liquids by an optical method. Proc. Phys. Soc. London 1923, 36, 313–335. (25) McBain, J. W.; Liu, T. H. Diffusion of electrolytes, nonelectrolytes and colloidal electrolytes. J. Am. Chem. Soc. 1931, 53, 59–74. (26) McBain, J. W.; Dawson, C. R. The diffusion of potassium chloride in aqueous solution. Proc. R. Soc. London A 1935, 148, 32–39. (27) Gordon, A. R. The diffusion constant of an electrolyte, and its relation to concentration. J. Chem. Phys. 1937, 5, 522–526. (28) Hartley, G. S.; Runnicles, D. F. The porous diaphragm method of measuring velocity, and the velocity of diffusion of potassium chloride in water. Proc. R. Soc. London A 1938, 168, 401–419. (29) Rysselberghe, P. V. Interpretation of diffusion data for some strong electrolytes. J. Am. Chem. Soc. 1938, 60, 2326–2331. (30) Vinograd, J. R.; McBain, J. W. Diffusion of electrolytes and of the ions in their mixtures. J. Am. Chem. Soc. 1941, 63, 2008–2015. (31) Harned, H. S.; French, D. M. A conductance method for the determination of the diffusion coefficients of electrolytes. Ann. N.Y. Acad. Sci. 1945, 46, 267–284. (32) Gordon, A. R. The diaphragm cell method of measuring diffusion. Ann. N.Y. Acad. Sci. 1945, 46, 285–308. (33) Harned, H. S.; Nuttall, R. L. The diffusion coefficient of potassium chloride in dilute aqueous solution. J. Am. Chem. Soc. 1947, 69, 736–740. (34) Harned, H. S.; Nuttall, R. L. The differential diffusion coefficient of potassium chloride in aqueous solution. J. Am. Chem. Soc. 1949, 71, 1460–1463. (35) Harned, H. S.; Nuttall, R. L. The diffusion coefficient of potassium chloride in aqueous solution at 25 °C. Ann. N.Y. Acad. Sci. 1949, 51, 781–788. (36) Harned, H. S.; Blake, C. A. The diffusion coefficient of potassium chloride in water at 4 °C. J. Am. Chem. Soc. 1950, 72, 2265–2266. (37) Gosting, L. J. A study of the diffusion of potassium chloride in water at 25° with the Gouy interference method. J. Am. Chem. Soc. 1950, 72, 4418–4422. (38) Stokes, R. H. An improved diaphragm-cell for diffusion studies, and some tests of the method. J. Am. Chem. Soc. 1950, 72, 763–767. (39) Stokes, R. H. Integral diffusion coefficients of eight uni-univalent electrolytes in aqueous solution at 25 °C. J. Am. Chem. Soc. 1950, 72, 2243–2247.

ARTICLE

(40) Stokes, R. H. Integral diffusion coefficients of potassium chloride solutions for calibration of diaphragm cells. J. Am. Chem. Soc. 1951, 73, 3527–3528. (41) Harned, H. S. Solutions of electrolytes. Annu. Rev. Phys. Chem. 1951, 2, 37–50. (42) Akeley, D. F.; Gosting, L. J. Studies of the diffusion of mixed solutes with the Gouy diffusiometer. J. Am. Chem. Soc. 1951, 73, 3527–3528. (43) Rulfs, C. Polarographic diffusion coefficients. J. Am. Chem. Soc. 1954, 76, 2071–2074. (44) Mills, R.; Adamson, A. W. The measurement of self-diffusion in electrolyte solutions. J. Am. Chem. Soc. 1955, 77, 3454–3458. (45) Longsworth, L. G. The temperature dependence of the Soret coefficient of aqueous potassium chloride. J. Phys. Chem. 1957, 61, 1557–1562. (46) Woolf, L. A.; Miller, D. G.; Gosting, L. J. Isothermal diffusion measurements on the system H2O-Glycine-KCl at 25°; Tests of the Onsager reciprocal relation. J. Am. Chem. Soc. 1962, 84, 317–331. (47) Firth, J. G.; Tyrrell, H. J. Diffusion coefficients for aqueous silver nitrate solutions at 25°, 35°, and 45° from diaphragm-cell measurements. J. Chem. Soc. 1962, 2042–2047. (48) McCall, D. W.; Douglass, D. C. The effect of ions on the self-diffusion of water. I. Concentration dependence. J. Phys. Chem. 1965, 69, 2001–2011. (49) Miller, D. G. Application of irreversible thermodynamics to electrolyte solutions. I. Determination of ionic transport coefficients lij for isothermal vector transport processes in binary electrolyte systems. J. Phys. Chem. 1966, 70, 2639–2659. (50) Woolf, L. A.; Tilley, J. F. Revised values of integral diffusion coefficients of potassium chloride solutions for the calibration of diaphragm cells. J. Phys. Chem. 1967, 71, 1962–1963. (51) Carman, P. C. Transport in concentrated solutions of 1:1 electrolytes. J. Phys. Chem. 1969, 73, 1095–1105. (52) Yeh, H. S.; Wills, G. B. Diffusion coefficient of sodium nitrate in aqueous solution at 25 °C as a function of concentration from 0.1 to 1.0 M. J. Chem. Eng. Data 1970, 15, 187–189. (53) Collings, A. F.; Hall, D. C.; Mills, R.; Woolf, L. A. A conductance-monitored diaphragm cell for diffusion measurements. J. Phys. E: Sci. Instrum. 1971, 4, 425–428. (54) Fell, C. J. D.; Hutchison, H. P. Diffusion coefficients for sodium and potassium chlorides in water at elevated temperatures. J. Chem. Eng. Data 1971, 16, 427–429. (55) Newman, J.; Chapman, T. W. Restricted diffusion in binary solutions. AIChE J. 1973, 19, 343–348. (56) Agar, J. N.; Lobo, V. M. M. Measurement of diffusion coefficients of electrolytes by a modified open-ended capillary method. J. Chem. Soc., Faraday Trans. 1 1975, 71, 1659–1666. (57) Tanaka, K. Effect of electrolytes on the self-diffusion coefficient of water. J. Chem. Soc., Faraday Trans. 1 1976, 72, 1121–1123. (58) Arkhipov, V. P.; Emel'yanov, M. I.; Samigullin, F. M.; Gaisin, N. K. Self-diffusion of water molecules in solutions of electrolytes at pressures up to 2500 kg/cm2. J. Struct. Chem. 1979, 19, 709–712. (59) Rard, J. A.; Miller, D. G. Mutual diffusion coefficients of BaCl2H2O and KClH2O at 25 °C from Rayleigh interferometry. J. Chem. Eng. Data 1980, 25, 211–215. (60) Chang, Y. C.; Myerson, A. S. The diffusivity of potassium chloride and sodium chloride in concentrated, saturated, and supersaturated aqueous solutions. AIChE J. 1985, 31, 890–894. (61) Myerson, A. S.; Chang, Y. C. Diffusional separation in ternary systems. AIChE J. 1986, 32, 1747–1749. (62) Pinto, N. G.; Graham, E. E. Multicomponent diffusion in concentrated electrolyte solutions: effect of salvation. AIChE J. 1987, 33, 436–443. (63) Leaist, D. G.; Hao, L. Gravitational stability of Taylor dispersion profiles. Revised diffusion coefficients for barium chloride-potassium chloride-water. J. Phys. Chem. 1993, 97, 1464–1469. (64) Lobo, V. M. M. Mutual diffusion coefficients in aqueous electrolyte solutions. Pure Appl. Chem. 1993, 65, 2613–2640. (65) Sami Selim, M.; Al-Naafa, M. A.; Jones, M. C. Brownian diffusion of hard spheres at finite concentrations. AIChE J. 1993, 39, 3–16. 4847

dx.doi.org/10.1021/je200808q |J. Chem. Eng. Data 2011, 56, 4840–4848

Journal of Chemical & Engineering Data

ARTICLE

(66) Al-Naafa, M. A.; Sami Selim, M. Measurement of diffusion coefficients by Taylor's method of hydrodynamic stability. AIChE J. 1994, 40, 1412–1417. (67) Lobo, V. M. M.; Ribeiro, A. C. F.; Andrade, S. G. C. S. Computation of the diffusion coefficients in aqueous electrolyte solutions from Onsager-Fuoss and Pikal theories. Bunsenges. Phys. Chem. 1995, 99, 713–720. (68) M€uller, K. J.; Hertz, H. G. A parameter as an indicator for waterwater association in solutions of strong electrolytes. J. Phys. Chem. 1996, 100, 1256–1265. (69) Bernard, O.; Cartailler, T.; Turq, P.; Blum, L. Mutual diffusion coefficients in electrolyte solutions. J. Mol. Liq. 1997, 73 (74), 403–411. (70) Leaist, D. G.; Lu, R. Diffusion coefficients measured by Taylor dispersion in a Fourier ring. Aqueous lanthanum chloride at 25 °C. J. Chem. Soc., Faraday Trans. 1997, 93, 1755–1761. (71) Lobo, V. M. M.; Ribeiro, A. C. F.; Verissimo, L. M. P. Diffusion coefficients in aqueous solutions of potassium chloride at high and low concentrations. J. Mol. Liq. 1998, 78, 139–149. (72) Ribeiro, A. C. F.; Lobo, V. M. M; Leaist, D. G.; Natividade, J. J. S.; Verissimo, L. P.; Barros, M. C. F.; Cabral, A. M. T. D. P. V. J. Solution Chem. 2005, 34, 1009–1016. (73) Lobo, V. M. M.; Ribeiro, A. C. F.; Valente, A. J. M. Diffusion in electrolytes and hydrogels. In International Workshop Transport in Fluid Multiphase Systems: From Experimental Data to Mechanistic Models, March 2930, 2004, RWTH Aachen, Aachen, Germany; pp 139. (74) Wakeham, W. A.; Nagashima, A.; Sengers, J. V., Eds. Measurement of the Transport Properties of Fluids; Blackwell Scientific Publications: Oxford, 1991. (75) Wakeham, W. A.; Assael, M. J.; Marmur, A.; De Coninck, J.; Blake, T. D.; Theron, S. A.; Zussman, E. Material properties: measurement and data. In Springer Handbook of Experimental Fluid Mechanics; Tropea, C., Yarin, A. L., Foss, J. F., Eds.; Springer-Verlag: Berlin, 2007. (76) Cussler, E. L. Diffusion: Mass Transfer in Fluid Systems, 3rd ed.; Cambridge University Press: New York, 2007. (77) Alizadeh, A.; Nieto de Castro, C. A.; Wakeham, W. A. The theory of the Taylor dispersion technique for liquid diffusivity measurements. Int. J. Thermophys. 1980, 1, 243–284. (78) Alizadeh, A. Binary diffusion coefficients for liquids. Ph.D. Thesis, Imperial College University of London, 1981. (79) Alizadeh, A. A.; Wakeham, W. A. Mutual diffusion coefficients for binary mixtures of normal alkanes. Int. J. Thermophys. 1982, 3, 307–323. (80) Levenspiel, O.; Smith, W. K. Notes on the diffusion-type model for the longitudinal mixing of fluids in flow. Chem. Eng. Sci. 1957, 6, 227–233. (81) Harris, K. R. On the use of the Edgeworth-Cramer Series to obtain diffusion coefficients from Taylor dispersion peaks. J. Solution Chem. 1991, 20, 595–606. (82) NIST. NIST Standard Reference Database Number 69 (http://webbook.nist.gov/chemistry/) (accessed July 4, 2011). (83) DETHERM 2007. DECHEMA (http://www.dechema.de/en/ detherm.html) (accessed July 4, 2011).

4848

dx.doi.org/10.1021/je200808q |J. Chem. Eng. Data 2011, 56, 4840–4848